MOTORTECH

The aim of the present project was to develop a preliminary intelligent system for the diagnosis of electric motors, in especial, large machines. Eventual failures in such machines can lead to severe consequences, not only in industry, where these motors take part in a wide variety of processes (most of them critical) and where their inspection and reparation may imply very significant costs, but also in other critical applications as power generation plants or electric traction systems.

The proposed system relied on the combination of different non-invasive techniques, i.e., that do not imply any interference with the usual operation of the machine. Nowadays, this is a crucial characteristic in many industrial processes. More specifically, the proposed system was based on the following techniques: analysis of machine currents (including both transient and stationary analysis), infrared thermography and stray flux analysis.

In the project, multiple simulations and experimental tests were developed in order to analyse the most suitable technique for the detection of each type of fault, as well as to determine its success rate. In this regard, all the possible faults that are likely to happen in electric motors were considered (namely, rotor failures, bearing damages, stator faults, ventilation system failures, misalignments, etc…), as well as the possible effects caused by external factors as the driven load or the supply.

The developed system integrated the previous techniques in advanced algorithms based on signal and image processing, pattern recognition and artificial intelligence tools, that were aimed to enable the automatic diagnostic of each fault, i.e., without requiring the intervention of expert users. The developed algorithms were validated in real motors, both at the laboratory and in industrial sites (especially, large motors).

Duration: January 2015-December 2018 Principal Investigator: Prof. JOSE A ANTONINO-DAVIU